What Should You Look for in AI Solutions for Procurement?
A Gartner report published in September 2018, entitled “The Impact of Artificial Intelligence on Procurement Software Applications” set out the reasons why many procurement leaders are hesitant about investing in artificial intelligence (AI) solutions. The headline reason can be summarized in one word: hype. Based on the Gartner research and our own extensive survey findings, this article discusses what procurement professionals are looking for, and what they should look for in future, to secure the benefits of artificial intelligence solutions in their organization. Because one thing is abundantly clear: those organizations that succeed in seeing through the hype and successfully implement solutions that drive value will reap significant competitive advantages.
The Stampede to Incorporate AI into Software Offerings
“Hype” tends to develop its own unstoppable momentum as vendors fight for attention. If you don’t say you’re incorporating disruptive technologies in software then, no matter how good your product and no matter how satisfied your customers, you’re not going to win the attention of the thought leaders and media commentators.
According to the Gartner report, “Artificial intelligence is rapidly emerging as the technology of choice for software application vendors seeking to gain differentiation through technology.” It quotes Gartner’s own Tech and Service Provider CEO Survey, which indicated that 75% intend to incorporate AI into their products by 2020, up from 39% in 2017. Gartner states that “The shift toward AI is natural, as previous investments such as cloud now provide a diminished competitive differentiation, while AI is attracting increasing demand especially for digital transformation.”
Seventy-five percent of application software vendor CEOs who participated in Gartner’s Tech and Service Provider CEO Survey indicated the intent to incorporate AI into their products by 2020. This is up from 39% in 2017.
And yet, procurement professionals are yet to be convinced. Gartner estimates the current penetration of the addressable market for AI to be somewhere between 1% and 5% and that procurement professionals are hesitant about investing in new software. Procurement people, we know, are especially hard-headed when it comes to new investment, and many are struggling to cope with stagnant or near-stagnant budgets. Chief Procurement Officers (CPOs) and other leaders are concerned about “difficulty in measuring business value and lack of clarity around the deployment scenario required for successful deployment of new software. These concerns result in procurement leaders hesitating to invest in new AI-powered procurement software applications.”
This finding squares with JAGGAER’s own research findings. Our recently published report, Digital Transformation in Procurement: How Close Are We? based on a global survey with responses from 391 CPOs and other senior procurement decision makers revealed that 52% of organizations have no plans to invest in AI, although those that did have plans could see its applicability across virtually all procurement processes. We also gauged which of six emerging technologies was most applicable overall. Artificial intelligence was the most popular choice among 19% of respondents, some way behind predictive analytics (28%). Significantly, the survey revealed that process efficiencies and cost optimization are regarded as the most important benefits of AI and similar technologies in the procurement function.
For the time being, the procurement function is focused on securing tangible benefits from digital transformation rather than achieving a longer-term strategic vision.
What is AI?
This leads us directly to the heart of the matter. Gartner recommends that software vendors “align their investments to high-priority and definable business cases that can be measured and demonstrate clear business value” and “prioritize deploying AI solutions that require simple deployment scenarios without much user training or process transformation to allay concerns about complexity or deployment.”
Before we do so, it is worthwhile stepping back here and asking the simple but fundamental question, “What is Artificial Intelligence” and perhaps more pertinently, “What is it not?” before proceeding to the crunch question: “what must happen as a prerequisite for securing concrete, measurable benefits from AI applications?”
The answer to the first question is, superficially, a simple one: AI is any technique that enables machines to mimic human behavior. But “human behavior” is pretty broad. Let’s take reading a book as an example. Reading consists of two fundamentally different human activities. First, there is the purely mechanical process of identifying patterns of letters as words, and sequences of words as sentences. This is what optical character recognition (OCR) does. OCR used to be regarded as a form of artificial intelligence, because it mimics this simple mechanical process. But the second human activity involved in reading is making sense of the text intellectually. OCR does not do this. Text mining and natural language processing (NLP) and natural language generation (NLG) do so, and are now rightly regarded as a form of AI, but while NLP can achieve certain very specific objectives, it falls way short of human intellect, which is capable of interpretation, initiative, responsiveness etc.
Therefore, let us distinguish between strong and weak AI. The aim of strong AI (also super intelligence or general AI) is to acquire or surpass the same intellectual skills as human beings. Strong AI does not just behave reactively, but also on its own initiative, intelligently and flexibly. As of today, it has not yet been possible to develop strong artificial intelligence.
Weak (or narrow) AI, by contrast, refers to systems that focus on solving specific application problems, based on the methods of mathematics and computer science, which are specially optimized for the relevant requirement. All currently existing systems (including the example given above, of NLP) fall within the category of weak artificial intelligence.
We can therefore conclude that any vendor that suggests that its AI solutions in some way displace human decision-makers is guilty of hype, and what’s worse, dangerous hype.
How Should You Set the Right Expectations of AI in Procurement Solutions?
We say “dangerous hype” because it feeds the misconceptions that inhibit digital transformation. Our JAGGAER survey indicated that around half of procurement professionals do not rate their knowledge of digital technologies as “up to date”. That has barely budged since 2017. Artificial intelligence (which is generally seen as the super intelligence referred to above) arouses suspicion in the popular imagination: fear that jobs will be lost, fear of the unknown and ethical questions, to which there are no simple answers. Procurement professionals are not isolated from these misconceptions.
So, let us draw back from the hype. What should procurement reasonably expect from AI solutions? Actually, there are many areas. They include the following areas:
AI can be successfully deployed to automate specific processes (or sub-processes) in, for example, spend management, contract management, audits and negotiation management.
Human knowledge can be captured and used to guide users of procurement software (customers, suppliers) to find answers to their questions.
Just like Amazon makes recommendations on purchases, procurement can deploy AI to make recommendations in areas such as supplier management, award decisions, category strategy etc.
AI can be and has been deployed to make accurate forecasts based on various scenarios in areas such as spend, risk management and supplier reliability.
Note that these are all limited in scope, but they all address common priorities among procurement leaders. We believe that the recommendations in this Gartner report highlight the priorities that vendors focus on as they continue to expand their product road map. However, this must involve dialog with customers and users. Customers also have a role to play. Not least because there is not, and never can be, a successful AI deployment without data. And not just data, but huge volumes of data.
For this reason, JAGGAER works very closely with all of the key stakeholders at customer organizations to identify not just high-priority use cases that align investment to business needs, but also the data sources that will be required to run them. We have developed solutions in all of the four categories above, and many of the sub-categories.
We will be accelerating this development over the coming years. JAGGAER’s whole innovation strategy and road map over the past couple of years has also been focused on building a single platform, JAGGAER ONE, which puts all procurement-related data on a single platform. This provides the essential fuel to drive AI. We will look at some specific examples in the next of this series of articles.
Gartner, The Impact of Artificial Intelligence on Procurement Software Applications, Balaji Abbabatulla, Magnus Bergfors, Patrick Connaughton, 14 September 2018
Gartner does not endorse any vendor, product or service depicted in its research publications, and does not advise technology users to select only those vendors with the highest ratings or other designation. Gartner research publications consist of the opinions of Gartner’s research organization and should not be construed as statements of fact. Gartner disclaims all warranties, expressed or implied, with respect to this research, including any warranties of merchantability or fitness for a particular purpose.